In Data Visualizations Related to Gas Prices, Tony Rose of Support Analytics Consulting shows a series of visualizations he’s culled from the web, all related to gasoline prices. One infographic stuck out, partly for its overuse of chart junk, and partly for Tony’s interpretation. Here is the original chart, which Tony took from www.thebiblog.com, followed by Tony’s comments, which I’ve condensed.
Here is a visual of what makes up the price for gas in 2004, when the average price was $1.85 per gallon versus 2008 where we are now paying roughly $4.03 per gallon on average. . . . One fact that I would have expected to see below is a dramatic increase in the distribution cost of gas between 2004 and 2008, which actually decreased. Seems like there should be an almost perfect correlation between distribution costs and the price of gas, right? Maybe the impact is hidden due to the category being both distribution and marketing.
I guess the chart colors and shading were chosen to remind us how grim the situation is. At first it struck me as odd that distribution costs went down (as it did Tony), but then I noticed that although crude tripled in price, its contribution to the total cost of a gallon of gas only increased by about half. I quickly converted the percentages to actual prices for each component in the price, and the story became more clear. I’ve normalized the 2008 price per gallon to $3.72, not $4.03 as stated in the chart above, to make the crude cost per gallon of gas proportional to the price per barrel of crude.
In 2008, all components except refining have increased in actual dollar contribution. Crude has increased by the greatest amount, but distribution and marketing is the second fastest growing component. The three charts below are (a) my version of the original chart of percentage contribution of the components of gasoline price, (b) my chart of the price contribution of these components, and (c) my chart of the increase in price of the individual components. I used horizontal bar charts so the labeling would work better.
(a) Percentage contribution to the price of a gallon of gasoline
(b) Cost contribution to the price of a gallon of gasoline
(c) Increase in costs of components of gasoline price
One chart doesn’t always tell the whole story, and sometimes the selected chart may not tell it as well as a different chart. If I had to squeeze one chart into the space allotted by my editor, I’d use the middle one, showing true cost inputs, not percentages. The percentage chart downplays the contribution of crude prices, and gives the false impression that all other inputs are decreasing in price.